ساختار بازار درون زا و رقابت در صنعت تولید آبمیوه آمریکا در قرن 19

This study examines market structure and competition in the brewing industry in American frontier states around 1880. Although comprehensive firm-level data are unavailable, available data provide information about the presence of firms in western frontier markets and characteristics of those markets. The analysis uses these data in a discrete dependent variable econometric model to estimate the size of a market needed to support a certain number of breweries and the impact of market characteristics on market structure. In addition, the estimated market sizes needed to sustain different numbers of firms provide information about the competitive effects of entry in the early brewing industry. The results indicate that the presence of second and third brewing firms significantly increased competition in a market implying that duopoly markets were not perfectly competitive. Moreover, these findings suggest that brewing firms were not engaged in collusive behavior.

A longstanding issue in industrial organization concerns the relationship between market structure and market outcomes. Many conventional models hold that prices tend to the competitive level as the number of firms grows, although the rate of convergence depends on numerous factors such as the elasticity of demand. On the other hand, the theory of contestable markets (Baumol et al., 1982) predicts that the threat of entry is sufficient to discipline incumbents so that the actual effect of additional firms will be minimal. Still other scenarios can be constructed in which, for example, firms are able to successfully maintain a cartel following entry, thereby preventing price–cost margins from falling with increases in the number of firms. Beginning with Bain (1956) and continuing through, for example, Weiss (1989), a large body of research has examined correlations between measures of market performance, such as prices or price–cost margins, and measures of market structure, such as concentration ratios or counts of active firms. However, this structure–conduct–performance approach has been subject to a number of criticisms. The frequent use of price–cost margins and other variables culled from accounting data raises doubts about the ability of these studies to accurately measure market performance. A more fundamental criticism lies in the endogeneity of market structure. An observed market structure is the outcome of the profit-maximizing decisions of numerous actual and potential firms. When making entry decisions, these agents take into account numerous market conditions including the severity of barriers to entry, the nature of post-entry price competition, and the optimal entry decisions of other firms in the market. Thus, the simultaneous causal link between structure, conduct and performance confounds attempts to draw clear conclusions from this line of research. A number of recent studies have developed techniques to address these problems.1 In particular, Bresnahan and Reiss (1990) and Berry (1992) construct and estimate discrete dependent variable econometric models that explicitly confront the endogeneity problem by addressing the relationship between market structure and the entry decisions of firms. Studies such as Reiss and Spiller, 1989, Stavins, 1995, Downes and Greenstein, 1996, Berry and Waldfogel, 1999 and Mazzeo, 1999 and Abraham et al. (2000) have employed variants of this approach to examine market structure in a range of different industries. Another line of research, surveyed in Bresnahan (1989), endeavors to draw inferences about competition in the absence of direct measures of market performance such as price–cost margins. Bresnahan and Reiss (1991) propose a link between models of endogenous market structure and those examining market competition. Lacking information on price–cost margins, they develop tests based on estimates of entry thresholds or, in other words, the market sizes required to support different numbers of firms. Variation in these breakeven market sizes over market structures provides information about how competition changes with entry of additional firms. In particular, breakeven market sizes per firm which do not vary over market structures indicate that additional entrants are not changing competitive conduct very much. In other words, if the duopoly and triopoly entry thresholds are 2000 and 3000, respectively, so that the per firm breakeven size is 1000 in both cases, then this result suggests that entry of the third firm does not greatly increase competition in the triopoly market. On the other hand, if additional entrants lead to more strenuous competition, then one would expect the triopoly entry threshold to exceed 3000 so that the triopoly breakeven market size per firm would be greater than 1000. This paper employs these techniques to examine the impact of entry and market structure on competition in the American brewing industry in the late 19th century. Specifically, this study investigates features of localized markets needed to support entry of breweries using data on the presence of brewing firms in frontier towns around 1880. Thus, the unit of analysis is a frontier town, and the issue of interest is the number of breweries present in that market around 1880. The inquiry focuses on this region and time period primarily for purposes of market definition to insure that the characteristics of markets and the entry decisions of firms are well defined. The key data requirements are information about the presence of firms in frontier markets and characteristics of those markets. I compile the former from American Breweries ( Bull et al., 1984), a directory of breweries active in the United States from approximately colonial times to the present, while the latter are readily available from historical censuses. Following Bresnahan and Reiss (1990) and Berry (1992), the study involves a model of entry which provides conditions characterizing firms’ entry decisions into individual markets. The empirical implementation uses these conditions in a discrete dependent variable model to estimate the impact of market characteristics on firm entry and to determine the size of markets needed to support a certain number of firms. As in Bresnahan and Reiss (1991), variation in these estimated market sizes over market structures permits inferences about the impact of additional firms on competition. The estimation results suggest that entry of second and third firms significantly influenced competition in a market. On the one hand, these findings suggest that brewing firms were not engaged in collusive behavior. On the other hand, the results indicate that duopoly markets were not perfectly competitive as the addition of a third firm did have an effect on competition in a market. Thus, while competition intensified with the entry of additional firms, the increase in competition was not severe enough to drive markets to perfectly competitive levels in the range of market structures which I observe. Beyond providing evidence about the competitive effects of entry in oligopoly markets, this paper also studies an interesting industry during an era that past work has not examined.2 Research on the brewing industry has generally focused on the post-Prohibition period (e.g. Greer, 1981; Tremblay, 1987; Baker and Bresnahan, 1988; McGahan, 1991; Hausman et al., 1994). Unlike the mature industries which previous empirical studies of entry and endogenous market structure examine, brewing in the 1800s was an emerging industry. Moreover, the local nature of the industry in that period permits clean definition of individual markets. The paper is organized as follows. Section 2 presents a version of recently developed techniques for endogenizing market structure and describes the inferences which entry thresholds permit about the impact of entry on market competition. Section 3 discusses the data and the American brewing industry in the latter part of the 19th century. Section 4 presents and interprets the estimation results while the closing section concludes with suggestions for future research.

This paper provides insight into some of the market conditions facing brewing firms in the late 1800s. The analysis uses data on the presence of breweries in western frontier markets and characteristics of those markets in a discrete dependent variable model to analyze the factors influencing market structure. The brewing industry at that time serves as an ideal application of such an approach since the existence of high transportation costs permits reasonable identification of local markets. The results suggest that the demographic composition of a market had a definite impact on entry as brewing was more profitable in markets which were more heavily German and male. Moreover, market growth likely had an important effect on the entry decisions of potential brewers. It was also found that a market of relatively small size, although perhaps not by frontier standards, was able to support entry of one firm while substantially larger towns were required for more firms to enter. More interestingly, this study provides evidence about the competitive effects of entry in oligopoly markets while taking into account the endogeneity of market structure and the inability to observe many economic variables. In the case of the brewing industry around 1880, this paper argues that entry of a second firm did lead to more strenuous competition as a market needed to more than double in size to support a duopoly rather than a monopoly. Similarly, markets became more competitive with the entry of third firms as triopoly market size had to increase disproportionately relative to duopoly market size. These results imply that brewing firms were not colluding as competition intensified with the entry of additional firms. However, having two firms in a market was not sufficient for perfect competition. Extension of the techniques which this paper employs could also serve to expand knowledge about the evolution of industries such as the brewing industry. In particular, an explicitly dynamic model along with panel data on the entry and exit decisions of firms, such as American Breweries, could provide additional insight into the technological, demand and competitive effects which influence entry and market structure. Horvath et al. (1999) study of the brewing industry provides a preliminary step in this direction. However, their theory about entry and exit patterns in the brewing industry, namely information accumulation leading to excessive entry which is then followed by mass exit, is only one possible explanation for events in this type of industry. More complete models could test such an explanation against those related to, for example, demand or cost factors. Such extensions could provide valuable insight into the underlying causes of market structure and the impact of structure on the performance of markets.